Publication Date


Degree Program

Department of Economics

Degree Type

Master of Arts


The purpose of this paper is to re-estimate the impact of Social Security on aggregate private saving behavior by using new time series data for the period 1947 through 1993. The analysis is based on the life-cycle hypothesis developed by Modigliani and Ando. Feldstein's extended life-cycle model, incorporating Social Security Wealth (SSW) into the traditional life-cycle model, provides the building block for this paper. A Generalized Least Squares (GLS) procedure is used to perform the time series analysis in order to avoid autocorrelation, which is usually associated with the Ordinary Least Squares (OLS) procedure. Compared with OLS estimations, the results from the GLS procedure do not indicate a large divergence regarding the statistical significance of the coefficient of SSW as well as those of other variables included except the smaller estimated parameter values. The major finding in this paper is that SSW does depress potential personal and private saving. The result support Feldstein's 1974 findings, but the magnitude of the coefficient of SSW is smaller. Feldstein's estimations, by using OLS, indicated that SSW reduced potential personal saving by 50% and induced total private saving to decrease by 38%. Our estimates imply that SSW reduces personal and private savings by 44% and 13% , respectively. The most important finding in this paper is that if SSW reduces total private saving by 13%), in the long run the decrease in the rate of private saving would also depress the private capital stock by 13%, which implies a substantial reduction in GDP and a lower level of real income.



Included in

Economics Commons